Re: [PATCH] Document .Nbyte assembler directives

["Tobin C. Harding" <me at tobin dot cc>]

On Mon, Mar 06, 2017 at 04:45:51PM +1100, Tobin C. Harding wrote:
> On Mon, Mar 06, 2017 at 12:18:40PM +1030, Alan Modra wrote:
> > On Mon, Mar 06, 2017 at 11:10:56AM +1100, Tobin C. Harding wrote:
> > > GAS contains directives .2byte .4byte and .8byte. These directives are
> > > used in the test suite but are undocumented.
> > 
> > However, not all versions of gas support these directives.  You should
> > state that these directives are for ELF targets, the added
> > documentation should be inside @ifset ELF @end ifset, and the menu
> > under node Pseudo Ops needs updating.
> 
> Version 2 implements above suggestions (and successfully builds).
> 
> > 
> > > Add documentation to GAS section 'Assembler Directives' for each of
> > > directive.
> > > 
> > > ---
> > > 
> > > I was not able to build work out how to build the docs locally. This
> > > is my first patch to binutils. Please advise if format of patch or
> > > patch content is not how the project likes it. Am happy to re-work to
> > > fit in with your standards.
> > 
> > It is always dangerous to post a patch you haven't tested, as
> > demonstrated by the unbalanced @ifclear in your patch..
> > 
> > You will need to install makeinfo and perl to build the docs, then
> > "make info".  (There are other targets if you want to make pdf, dvi
> > etc. docs.  See gas/doc/Makefile.)
> 
> Thanks for your response, bad form of me to have done this. With your
> direction I was able to build the docs. I was not able to get the test
> suite running though. Since I am not across the source I am unsure if
> the documents I have written for .8byte are correct. I am a bit out of
> my league documenting things I don't understand but if you have the
> time (and patience) to please point me in the right direction I will
> try to verify it before submitting version 2. So far I am under the
> impression .8byte and .quad are alias'. Surely .8byte cannot accept 16
> bytes however, like .quad can if bignum-16 is set.

Please ignore this. I have verified that the behaviour is as the patch
documents. Sorry for the noise.

thanks,
Tobin.